You may find something useful here:
- https://andrewpwheeler.com/2017/06/12/testing-the-equality-of-coefficients-same-independent-different-dependent-variables/
HTH.
Hi Samuel Oluwaseun Adeyemo. What version of SPSS do you have? I ask, because I see no option for SUR in my version 28, and a search for "seemingly unrelated regression" in the documentation for SPSS 29 (the most recent version) comes up empty:
https://www.ibm.com/docs/en/search/%22seemingly%20unrelated%20regression%22?scope=SSLVMB_29.0.0
That makes me wonder if you are using an extension command. I cannot find a relevant extension command via Extensions > Extension Hub, however. So I hope you can clarify how you are able to find SUR in your Regression menu. Thanks.
Samuel Oluwaseun Adeyemo , please confirm that you wrote that yourself. I placed the question into ChatGPT, which I think everyone knows often makes up things to try to sound useful, and got something very similar the answer that you give.
To estimate Seemingly Unrelated Regression (SUR) using SPSS, you can follow these steps:
Remember, SUR assumes that the error terms across different equations are correlated, and this correlation needs to be taken into account while estimating the coefficients. The SPSS output will provide you with estimates that account for this correlation, so make sure to carefully examine the output to interpret the results.
Bruce Weaver , we were looking at how it did with how to questions for R, and they had that same style but for most of our questions it produced the wrong answer. Might be worth checking his other answers to see if there is a pattern.
Daniel Wright, I suspect that Samuel Oluwaseun Adeyemo may have posted another response from ChatGPT in this thread, where he cited an article that cannot be found on the journal website:
https://www.researchgate.net/post/In_linear_regression_does_the_dependent_variable_have_to_be_normally_distributed
I am curious about Samuel Oluwaseun Adeyemo motivations. Perhaps he can answer. If using ChatGPT he is presumably well aware that it makes up stuff to sound convincing. This suggests, if true, that he is deliberately trying to misinform other people. But, before passing judgment, it is worth hearing his side of this.
Samuel Oluwaseun Adeyemo, in your April 28 post in this thread, you advised the OP to do this in SPSS:
Access the "Regression" menu:From the menu bar, go to "Analyze" > "Regression" > "Seemingly Unrelated Regression".
As I said in an earlier reply, I see no SUR command in my copy of SPSS 28, and I can find no evidence that SUR was added in SPSS 29, the current version. This makes me very curious about why you say you can see it. You could clear everything up by explaining that. I would like to know how to gain access to SUR if it is indeed available in SPSS. Thank you for clarifying.
Bruce Weaver , he also states that SPSS has a procedure for SUR with correlated errors. Given the specifics that Samuel Oluwaseun Adeyemo 's post includes either he is working with a system beyond what (https://www.ibm.com/support/pages/seemingly-unrelated-regression) is refers to that we don't know about, or there is some other explanation for how he came up with this information. Unfortunately in his response above he did not answer where he got this information. Since he didn't answer on the other thread, the "using a new version" explanation seems less likely. To directly answer his question: Yes, there are issues. The details you provide do not appear accurate for the versions of SPSS that Bruce and I have and counter what IBM said a couple of years ago. So, paraphrasing what Bruce said originally, if something new exists in SPSS it would be good if you could tell people that it is new and give a source for them to learn more. If something new does not exist, then you should explain why you told people it exists.
Chuck A Arize, I don't know which AI chat-bot you used to generate that response, but had you read any of the earlier posts in this thread you would have known someone already posted this same advice and that it is wrong. There is no SUR option under Analyze > Regression in the SPSS GUI. You might also have noticed that this page on Andrew Wheeler's site has a possibly relevant example using the GLM command:
- https://andrewpwheeler.com/2017/06/12/testing-the-equality-of-coefficients-same-independent-different-dependent-variables/
The relevant GUI dialog is found by clicking on Analyze > General Linear Model > Multivariate. Please inform your chat-bot.
Bruce Weaver , do you think the account for " Chuck A Arize " is of a real person? It is difficult to imagine a real person doing this.
" Chuck A Arize " even gave a response to the discussion that was created by a Chatbot, with what is an LLM checker said was AI created.
I wondered the same thing, Daniel Wright, and I have just written to Dr. Arize at his university email address to ask if he has really posted 90 messages (so far) today (https://www.researchgate.net/profile/Chuck-Arize/answers). It could well be that his RG account has been hacked without his knowledge.
- Badges
- Science method
- Similar topics
- Teaching